Schema validator: Check your schema before Google does

This free schema validator reads every JSON-LD, Microdata and RDFa item on a page, checks it against all 922 schema.org types and 1,538 properties, and shows which Google rich results it can get, what stops the rest, and the fix to paste. It also shows what AI crawlers get, since they run no JavaScript.

How the schema validator works

The schema validator works in four steps, on our server, with no sign-up and nothing kept.

Read the page. Type a page's address and our server reads it as a browser does, without running JavaScript. Or paste the code: one JSON-LD object, a script tag, or a whole page's source.
Find every item. It reads each JSON-LD script with a strict parser that says on which line and column a script breaks, and repairs the common mistakes so you can copy the fixed code. It reads Microdata and schema.org RDFa too.
Check the words and the values. Every type and property is checked against schema.org's vocabulary, version 30.1, and every value against what its property takes: a date, a number, a URL or one of a list of values.
Check Google's rules and AI crawlers. Items are matched to the rich results Google shows today, with what each one needs. Then the page is asked for again as GPTBot, to see whether AI crawlers get the same schema.

What the schema checker tests

The schema checker tests four things, and sorts what it finds into errors, which stop search engines from reading a part, warnings, which make it weaker, and notes.

The JSON-LD syntax

A missing comma, a comma after the last item, curly quotes from a word processor or a comment makes a whole JSON-LD script unreadable. The checker names the line and column, says what is wrong in plain words, and hands back the repaired script.

The schema.org types and properties

Types and properties are case-sensitive: "Brand" is a type, "brand" is the property. The checker flags words schema.org doesn't have, suggests the one you meant, and names properties that don't belong to their item's type, like author on a Product.

The values fit

Dates must be ISO 8601, like 2026-10-22. Prices are plain numbers with the currency in priceCurrency. Availability is a value like https://schema.org/InStock. A nested item should be one of the types its property expects.

Required properties for rich results

For each item, the checker lists the Google rich results it can get, like Product snippet, Merchant listing and Review snippet, the required properties that are missing, such as a review's author, the ones Google recommends, and the rich results Google has stopped showing, like How-to.

Schema markup validator or Rich Results Test?

The Schema Markup Validator and Google's Rich Results Test answer different questions, and this checker answers both, plus one neither asks.

Question	Schema Markup Validator	Rich Results Test	This checker
Is every type and property schema.org's?	Schema Markup Validator: Yes	Rich Results Test: Only for the items it supports	This checker: Yes, with the word you meant
Which Google rich results can it get?	Schema Markup Validator: No	Rich Results Test: Yes	This checker: Yes, and the retired ones
Where does a broken script break?	Schema Markup Validator: It shows the error	Rich Results Test: It shows the error	This checker: Line, column, and the repaired script
What do AI crawlers get?	Schema Markup Validator: No	Rich Results Test: No: it renders the page as Google does	This checker: Yes: the page as GPTBot gets it

The Schema Markup Validator at validator.schema.org took over from Google's Structured Data Testing Tool. For Google's own final word, run a page through the Rich Results Test too.

Common schema errors and their fixes

Common schema errors come from a handful of mistakes, and each one has a fix you can make in a minute.

A comma after the last item. JSON allows no comma before a closing brace or bracket. Delete it.
Curly quotes. Code pasted through a word processor gets “curly” quotes. JSON needs straight ones (").
No @context. Without "@context": "https://schema.org", nothing says the words are schema.org's.
The wrong case. "@Type", "Name" and "Brand" as a property are skipped. Write @type, name and brand.
A property on the wrong type. price belongs in an Offer, not on the Product. ratingValue belongs in an AggregateRating.
Prices with signs. "$1,299.99" is text. Write "1299.99" and "priceCurrency": "USD".
Dates written out. "October 5, 2026" isn't read as a date. Write 2026-10-05.
Schema added by JavaScript. Schema put on the page by Google Tag Manager or a script reaches Google after rendering, but AI crawlers that run no JavaScript never see it. Put it in the HTML the server sends.

Does schema help AI search?

Schema helps AI search when it states facts a crawler can confirm on the page, not as a ranking trick: no schema type ranks a page by itself.

What it does well

Schema hands machines facts in a form they can't misread: your name and the names you go by, your logo, your profiles in sameAs, a product's price. The organization schema generator writes that part for every name you go by.

What it can't do

Schema that says something the page doesn't is ignored, or worse. Google's rules ask that markup match what people see, and an AI answer quotes your text, not your JSON. Put every fact in a plain sentence too, the way the About Us page checker asks.

Crawlers that run no JavaScript

GPTBot, ClaudeBot and PerplexityBot read the HTML your server sends. Schema written into that HTML reaches them; schema added later by a script doesn't. The AI visibility checker tests whether those crawlers can reach the page at all.

The rest of the head

Shared links show your Open Graph tags, not your schema. To see the title, description and image each app shows, run the page through the Open Graph checker.

Schema validator questions

These schema validator questions come up most: what it reads, how it differs from Google's tools, and what happens to your code.

Is the schema validator free?

Yes. The schema validator is free with no sign-up: 60 pages and 300 pasted checks a day, more when you sign in. It makes no paid calls, so there is nothing to sell you at the end.

Which formats does it read?

It reads JSON-LD, the format Google recommends, plus Microdata and schema.org RDFa, all on one page. It reads schema.org's vocabulary; items in other vocabularies are named and left out.

Does it run JavaScript?

No. It reads the HTML a page sends, the way AI crawlers like GPTBot and ClaudeBot do. Schema added by a script or Google Tag Manager won't show here, and those crawlers can't see it either. Google renders JavaScript, so its Rich Results Test can show more.

Valid schema, but no rich result in Google. Why?

Valid schema makes a page eligible, and Google still decides what to show. It weighs the page's quality, whether the markup matches what people see, and the search. Some rich results show only in some countries or on desktop: breadcrumbs, for one, no longer show on phones.

What is the difference between an error and a warning?

An error means search engines skip that part: a script that can't be read, a word schema.org doesn't have, a property on the wrong type, or a property Google needs for a rich result. A warning means it is read but weaker, like a date in the wrong format or a property Google recommends.

Why is a property not recognized for my type?

Each schema.org property belongs to certain types. price belongs to Offer, so on a Product it is skipped: put it in the Product's offers. The checker names the types each property belongs to.

Is FAQ or How-to schema still worth adding?

Not for Google's rich results: Google stopped showing How-to results in 2023 and FAQ results for every site on May 7, 2026. The markup is still valid schema.org, so it does no harm, and other search engines and AI tools can still read it.

Do you keep my code or the pages I check?

No. The page or the code is read for the check and thrown away. Nothing is stored, and the result lives only in your browser tab.